Practical Use for AI and Machine Learning in Science
AI is practically everywhere right now. But what practical use does it have in science? How can machine learning and large language models help in Astronomy and Cosmology? What can they look like and what use can we get out of them?
🟣 Guest: Dr Mike Smith https://www.mjjsmith.com/
Watch the video here (with no ads) or on YouTube: https://youtu.be/SexDIF__8l4
Practical Use for AI and Machine Learning in Science (Vimeo)
Comments (4)
Thanks for your responses They give me hope for the future of science
AI is very useful especially when trying to seek out patterns within a large dataset. This is akin to what Fraser indicated that AI can do a great job finding needles in a haystack. Humans may not be able to see what AI can point out as a fresh perspective on an issue buried deep in a bunch of data. Science can tolerate ideas and observations that may be off and it is the community that will eventually either verify the observations or disprove it through community independent observations and community consensus. Trust but verify anything presented especially by AI. I do this with chatGPT when I query anything. I use other sources of information before I totally buy into an answer from this tool.
To me it seems there are two conflicting drivers of AI. Veracity of answers verses efficiency of derivation. Due to short attention span of humans the latter may prevail and AI produce answers that look good but are wrong.
I am concerned that the output of the astronomy AI will not be verified at sensible by many people Similar to GPS where people follow it blindly and end up lost How will the science community audit the conclusions made with AI?